This copy is for personal use only. To order printed copies, contact reprints@rsna.

org
Imaging

Supplemental Screening Breast

US in Women with Negative
Original Research  n  Breast

Mammographic Findings:
Effect of Routine Axillary Scanning1
Su Hyun Lee, MD
Purpose: To evaluate the effect of routine axillary scanning when
Ann Yi, MD
supplemental screening breast ultrasonography (US)
Myoung-jin Jang, PhD
is performed in women with negative mammographic
Jung Min Chang, MD findings.
Nariya Cho, MD
Woo Kyung Moon, MD Materials and This retrospective review included 12 844 screening breast
Methods: US examinations performed in 8664 asymptomatic women
aged 40 years or older with dense breasts and negative
results for cancer at mammography performed between
January 2012 and December 2014. Bilateral whole-breast
US was performed with a handheld device by one of 10 ex-
perienced radiologists. The bilateral axillae were routinely
scanned, and representative images were documented in
all examinations. The abnormal interpretation rate (AIR),
cancer detection rate (CDR), and positive predictive value
(PPV) of screening breast US with and without axillary
scanning were calculated. The 95% confidence intervals
(CIs) were calculated for cancer detection after an abnor-
mal finding at screening US.

Results:
1
 From the Department of Radiology (S.H.L., J.M.C.,
N.C., W.K.M.) and Medical Research Collaborating
Center (M.J.J.), Seoul National University Hospital, 101
Daehak-ro, Jongno-gu, Seoul 03080, Republic of Korea;
and Department of Radiology, Seoul National University
Conclusion:
Hospital Healthcare System Gangnam Center, Seoul, Korea
(A.Y.). Received May 27, 2017; revision requested June 29;
revision received July 15; accepted August 10; final version
accepted August 24. Address correspondence to W.K.M.
(e-mail: moonwk@snu.ac.kr).
Supported by Korean Society of Breast Imaging & Korean
Society for Breast Screening (KSBI&KSFBS-2013-02).
The frequency of positive axillary findings was 3.5 per 1000
(14 of 4009) baseline screening US examinations and 2.2
per 1000 (19 of 8835) subsequent screening US examina-
tions. Of the 33 women with 33 positive axillary findings,
11 had positive breast findings; none were diagnosed with
breast cancer. The remaining 22 women showed positive
findings only in the axilla. The axillary findings revealed
no malignancy at biopsy (n = 12) or during 22–54-month
follow-up (n = 21) (95% CI: 0%, 10.6%). Without rou-
tine axillary scanning, the AIR of screening US decreased
from 15.2% (610 of 4009 examinations) to 15.0% (602 of
4009 examinations) at baseline US and from 8.1% (714 of
8835 examinations) to 7.9% (700 of 8835 examinations)
at subsequent US examinations, and the PPV for biopsy
performed increased from 6.0% (five of 83 examinations)
to 6.4% (five of 78 examinations) at baseline US and from
7.6% (13 of 170 examinations) to 7.9% (13 of 164 ex-
aminations) at subsequent US examinations, without a
change in the CDR.
Routine axillary scanning during screening breast US had
no effect on additional cancer detection, but rather in-
creased the number of false-positive results. However, the
conclusions based on these findings must be tempered by
the low rate of positive findings.
q
 RSNA, 2017

q
 RSNA, 2017

830 radiology.rsna.org  n  Radiology: Volume 286: Number 3—March 2018

BREAST IMAGING: Effect of Routine Axillary US in Women with Negative Mammographic Findings Lee et al
M
ammography is the only
screening method that has
been proven in randomized
controlled trials to reduce breast can-
cer–related mortality (1,2). However,
the sensitivity of mammography has
been reported to be as low as 30%–
48% in women with dense breasts
(3,4) and the risk of developing breast
cancer is higher in women with dense
breast tissue than in those without
dense breast tissue (5,6). Breast ul-
trasonography (US) and tomosynthesis
are the most commonly used supple-
mental screening modalities in women
with dense breasts and no other risk
Advances in Knowledge
nn The frequency of positive axillary
findings at screening US in
women with negative mammo-
graphic findings was 3.5 per 1000
(14 of 4009) baseline US exami-
nations and 2.2 per 1000 (19 of
8835) subsequent US examina-
tions; biopsy (n = 12) and at
least 22 months of follow-up (n =
21) for the 33 positive axillary
findings in 33 women revealed no
malignancy (95% confidence
interval [CI]: 0%, 10.6%).
nn Supplemental screening breast
US depicted 1.5 cancers (95%
CI: 0.5, 3.3) per 1000 baseline
examinations and 2.2 (95% CI:
1.4, 3.4) per 1000 subsequent
examinations with or without
axillary scanning; without routine
axillary scanning, the abnormal
interpretation rate of screening
US decreased from 15.2% (610
of 4009 examinations) to 15.0%
(602 of 4009 examinations) at
baseline US and from 8.1% (714
of 8835 examinations) to 7.9%
(700 of 8835 examinations) at
subsequent US examinations, and
the positive predictive value for
biopsy performed increased from
6.0% (five of 83 examinations) to
6.4% (five of 78 examinations) at
baseline US and from 7.6% (13
of 170 examinations) to 7.9% (13
of 164 examinations) at subse-
quent US examinations.
Radiology: Volume 286: Number 3—March 2018  n  radiology.rsna.org 831
factors (7). Breast US has many advan-
tages in that it is widely available, is
well tolerated by women, and requires
no ionizing radiation. Supplemental
screening breast US can depict small
mammographically occult cancers and
reduce the number of interval cancers
(8–15). In most previous studies, hand-
held US was used for breast screening.
However, implementation of handheld
US as a screening tool remains contro-
versial because of its false-positive rate,
the variability among operators, and the
considerable physician time required
for image acquisition (7). Recently,
three-dimensional automated breast US
is being investigated as a potential so-
lution for women with heterogeneously
and extremely dense breasts. Some
studies (16,17) have demonstrated sig-
nificantly improved cancer detection
with the addition of automated breast
US to screening mammography.
The axillary region can be included
when breast US is performed with a
handheld device. Scanning of the axilla
is considered optional because it may in-
crease scanning time and result in more
false-positive findings (18,19). How-
ever, the bilateral axillae are routinely
scanned and representative images are
documented in many institutions when
handheld breast US is performed be-
cause breast cancers have been known
to manifest as an isolated axillary nodal
metastasis without any clinically or ra-
diologically detectable breast tumors
(20), and various lesions, both benign
and malignant, can develop in the ac-
cessory breast tissue, which is present
Implications for Patient Care
nn Routine axillary scanning during
screening breast US does not
provide additional breast cancer
detection, but rather increases
the number of false-positive
results leading to recall examina-
tions and biopsies.
nn Additional axillary US may not be
necessary when automated
breast US or breast US with a
handheld device is performed in
women with negative findings at
screening mammography.
in 0.6%–6% of the general population
(21). In addition, scanning of the axilla
reinforces the importance of scanning
of the posterior breast when perform-
ing breast US with a handheld device
(19). Conversely, axillary regions can-
not be covered with automated breast
US owing to its wide field-of-view trans-
ducer, so when automated breast US
is performed, questions remain as to
whether additional axillary US should
be included. Yet despite these consid-
erations, relatively little is known about
the effects of routine bilateral axillary
scanning during screening breast US.
Therefore, the purpose of this study
was to evaluate the effect of routine
axillary scanning during supplemental
screening breast US in women with
negative mammographic findings.
Materials and Methods
Study Population
This study was approved by our insti-
tutional review board, and the require-
ment to obtain informed consent was
waived. A search of our institutional
database identified 23 846 consecutive
https://doi.org/10.1148/radiol.2017171218
Content codes:
Radiology 2018; 286:830–837
Abbreviations:
AIR = abnormal interpretation rate
BI-RADS = Breast Imaging Reporting and Data System
CDR = cancer detection rate
CI = confidence interval
PPV = positive predictive value
PPV1 = PPV for abnormal interpretation
PPV2 = PPV for biopsy recommendation
PPV3 = PPV for biopsy performed
Author contributions:
Guarantors of integrity of entire study, S.H.L., W.K.M.;
study concepts/study design or data acquisition or data
analysis/interpretation, all authors; manuscript drafting or
manuscript revision for important intellectual content, all
authors; approval of final version of submitted manuscript,
all authors; agrees to ensure any questions related to the
work are appropriately resolved, all authors; literature
research, S.H.L., W.K.M.; clinical studies, S.H.L., A.Y.,
J.M.C., N.C., W.K.M.; statistical analysis, S.H.L., M.J.J.; and
manuscript editing, S.H.L., J.M.C., W.K.M.
Conflicts of interest are listed at the end of this article.
BREAST IMAGING: Effect of Routine Axillary US in Women with Negative Mammographic Findings Lee et al
Figure 1
Figure 1:  Study flow diagram.
screening mammography examinations
performed in 15  608 asymptomatic
women aged 40 years or older at Seoul
National University Hospital Healthcare
System Gangnam Center from January
2012 to December 2014. Among them,
15 158 screening mammography exam-
inations showed heterogeneously or
extremely dense breast tissue and neg-
ative results for cancer. Supplemental
screening with bilateral whole-breast
US was performed within 6 months
after mammography in 12 844 (84.7%)
examinations. We retrospectively re-
viewed results from all 12 844 screen-
ing US examinations in 8664 women;
4009 (31.2%) were baseline US exami-
nations and 8835 (68.8%) were subse-
quent US examinations (Fig 1).
Screening Mammography
Mammography was performed with
a dedicated digital mammography
unit (Senographe 2000D; GE Medical
Systems, Milwaukee, Wis). Standard
craniocaudal and mediolateral oblique
832 radiology.rsna.org  n Radiology: Volume 286: Number 3—March 2018
views were routinely obtained. Screen-
ing mammograms were interpreted by
one of 10 board-certified and breast fel-
lowship–trained radiologists with 3–20
years of experience (including S.H.L.,
A.Y., J.M.C., N.C., and W.K.M.). Mam-
mographic breast density was classified
as grade a (almost entirely fat), b (scat-
tered fibroglandular densities), c (het-
erogeneously dense), or d (extremely
dense) according to the American Col-
lege of Radiology Breast Imaging Re-
porting and Data System (BI-RADS) lex-
icon (22). Women with initial BI-RADS
assessment category 1 (negative) or 2
(benign) at screening mammography
and with heterogeneously dense (grade
c) or extremely dense (grade d) breast
tissue were assigned to undergo supple-
mental screening breast US.
Screening Breast US
Screening breast US was performed on
the same day as mammography in 12 779
of 12 844 examinations (99.5%) and
after a median of 26 days after
mammography (range, 1–181 days) in
the remaining examinations. Screening
breast US was performed and inter-
preted by one of 10 dedicated breast ra-
diologists with 3–20 years of experience
in breast imaging (including S.H.L.,
A.Y., J.M.C., N.C., and W.K.M.) and
with knowledge of the mammographic
results and clinical information by us-
ing a standardized technique (13). Bi-
lateral whole-breast US examinations
were performed with a handheld device
(EUB-8500, Hitachi Medical, Tokyo,
Japan; Acuson S2000, Siemens Med-
ical Solutions, Mountain View, Calif)
equipped with a 6–14- or 6–18-MHz
linear-array transducer. The bilateral
axillae were routinely scanned, and
representative images from all exami-
nations were documented. If the axil-
lary US image showed benign-appear-
ing lymph nodes only, a single view of
a representative axillary lymph node
was obtained. If there were positive
axillary findings assessed as BI-RADS
category 3 or higher, two orthogo-
nal view images of the axillary finding
were acquired. Time to perform ex-
aminations was not recorded, but in
our experience it takes approximately
5–20 minutes per patient to perform
bilateral breast and axillary scanning.
BI-RADS category 1, 2, 3, 4A, 4B, 4C,
or 5 was recorded for each breast and
axillary finding. Axillary lymphadenop-
athy with a round shape, cortical thick-
ness greater than 3 mm, or loss of fatty
hilum detected in women without a his-
tory of rheumatoid arthritis or systemic
lupus erythematosus was assessed as
BI-RADS category 4 (biopsy recom-
mendation) or 3 (short-term follow-up
recommendation) at the radiologist’s
discretion (23,24).
Data Collection
Demographic and imaging data were
obtained from a prospectively main-
tained institutional database. For pos-
itive screening US examinations, de-
fined as BI-RADS category 3, 4A, 4B,
4C, or 5, the reports were reviewed to
determine whether the positive find-
ings were in the breast, the axilla, or
both. For negative screening examina-
tions, the results for the breast and
BREAST IMAGING: Effect of Routine Axillary US in Women with Negative Mammographic Findings Lee et al

axilla were recorded as negative. Our Table 1

institutional pathology database was
searched for biopsy or surgery per- Demographic Characteristics of Women with and without Supplemental Screening US
formed within 1 year of the screening after Negative Mammographic Findings
US examination in all women, and the With Supplemental Without Supplemental
results were reviewed. Primary breast Screening US Screening US
cancers diagnosed within 1 year were Characteristic (n = 12 844) (n = 2314) P Value
identified and classified as invasive
Mean age 6 SD (y) 51.2 6 7.2 50.0 6 7.9 ,.001
or ductal carcinoma in situ. Cancers
Age group (y) … … ,.001
containing mixed invasive and in situ
 40–49 5809 (45.2) 1269 (54.8) …
components were classified as inva-
 50–59 5449 (42.4) 770 (33.3) …
sive. Women with only lobular car-
  60 1586 (12.3) 275 (11.9) …
cinoma in situ were not considered Family history of breast cancer* … … ,.001
to have cancer. For all women with  No 12 411 (96.6) 2282 (98.6) …
positive screening US results in the  Yes 433 (3.4) 32 (1.4) …
breast and/or axilla, an additional Personal history of breast cancer … … .459
medical record review was performed  No 12 732 (99.1) 2298 (99.3) …
to determine clinical outcomes during  Yes 112 (0.9) 16 (0.7) …
the follow-up period. If no tissue di- Breast density … … .037
agnosis of cancer was made within 1   Heterogeneously dense 9169 (71.4) 1701 (73.5) …
year, the result was considered to be   Extremely dense 3675 (28.6) 613 (26.5) …
disease negative. Type of screening US … … NA
 Baseline 4009 (31.2) NA …
Outcome Measurements and Statistical  Subsequent 8835 (68.8) NA …
Analysis
Note.—Except where indicated, data are numbers of examinations, with percentages in parentheses. NA = not applicable,
All analyses were conducted by using SD = standard deviation.
the screening examination as the unit * Family history of breast cancer in first-degree relatives.
of analysis; women may have under-
gone more than one breast US exam-
ination during the study period. Dif-
ferences in demographics and breast Statistical analyses were performed Axillary Findings at Screening US in
density between women who under- with software (SAS for Windows, ver- Women with Negative Mammographic
went screening US and those who did sion 9.3; SAS Institute, Cary, NC). Findings
not undergo screening US were as- The frequency of positive findings in
sessed by using the x2 test or Fisher Results the axilla at supplemental screening US
exact test for categorical variables and was 3.5 per 1000 (14 of 4009) baseline
the t test for continuous variables. The Demographics of the Study Population US examinations and 2.2 per 1000 (19
abnormal interpretation rate (AIR), The mean age (6standard deviation) of 8835) subsequent US examinations
the cancer detection rate (CDR), and of women at the time of screening US (Table 2). Of the 33 positive axillary
the positive predictive value (PPV) for was 51.2 years 6 7.2 (Table 1). Of the findings in 33 women (median age, 51
abnormal interpretation (PPV1), PPV 12 844 women, 12 411 (96.6%) had no years; age range, 41–77 years), 23 were
for biopsy recommendation (PPV2), family history of breast cancer in first- unilateral lymph node enlargements,
and PPV for biopsy performed (PPV3) degree relatives and 12  732 (99.1%) six were bilateral lymph node enlarge-
were calculated separately for baseline had no personal history of breast can- ments, three were soft-tissue masses in
and subsequent US examinations ac- cer. Mammographic breast density was the axillary fossa, and one was a mass
cording to BI-RADS guidelines (25). classified as heterogeneously dense in the accessory breast tissue (Table 3).
Exact 95% confidence intervals (CIs) (BI-RADS grade c) in 9169 of the 12 844 Of 33 women with abnormal axillary
were calculated for AIR and CDR. To women (71.4%) and as extremely dense findings, nine patients had previously
determine the hypothetical perfor- (BI-RADS grade d) in 3675 (28.6%). undergone one US examination with neg-
mance of screening breast US without Women who underwent supplemen- ative findings, and one patient had pre-
routine axillary scanning, the outcome tal screening breast US after negative viously undergone two US examinations
measures were calculated in two ways: mammographic findings had a higher with negative findings. The remaining 23
with combined data for the breast and mean age and higher incidence of fam- women underwent screening US exami-
axilla and with data for the breast ily history of breast cancer than did nation once during the study period.
only. Two-sided P , .05 was indicative women who did not undergo screening There was no known current malig-
of a statistically significant difference. breast US (P , .001 for both). nancy in the 33 women with abnormal

Radiology: Volume 286: Number 3—March 2018  n  radiology.rsna.org 833

BREAST IMAGING: Effect of Routine Axillary US in Women with Negative Mammographic Findings Lee et al

Table 2 soft-tissue masses and not concordant

in one woman with unilateral axillary
Results of Screening US for Breast and Axilla in Combination or Alone lymphadenopathy that had disappeared
Finding Breast and Axilla Breast Axilla at follow-up US after 25 months. US-
guided fine-needle aspiration biopsy was
Baseline screening US (n = 4009) performed for unilateral axillary lymph-
  Negative (BI-RADS 1, 2) 3399 (0) 3407 (0) 3995 (6) adenopathy in one woman, and the re-
 Positive 610 (6) 602 (6) 14 (0)
sult was negative for malignant cells.
  BI-RADS 3 501 (1) 499 (1) 8 (0)
Of the 20 axillary lesions assessed as
  BI-RADS 4A 105 (4) 99 (4) 6 (0)
BI-RADS category 3, 11 remained sta-
  BI-RADS 4B 3 (1) 3 (1) 0 (0)
ble or had disappeared at follow-up US
  BI-RADS 4C 1 (0) 1 (0) 0 (0)
(median 32 months; range, 22–48
  BI-RADS 5 0 (0) 0 (0) 0 (0)
Subsequent screening US (n = 8835)
months). The other nine axillary lesions
  Negative (BI-RADS 1, 2) 8121 (3) 8135 (3) 8816 (22) were considered benign because there
 Positive 714 (19) 700 (19) 19 (0) was no evidence of malignancy in the
  BI-RADS 3 527 (6) 519 (6) 12 (0) nine affected women after a median clini-
  BI-RADS 4A 179 (11) 173 (11) 7 (0) cal follow-up of 32 months (range, 26–54
  BI-RADS 4B 6 (0) 6 (0) 0 (0) months). The rate of cancer among those
  BI-RADS 4C 1 (1) 1 (1) 0 (0) with positive axillary findings was 0%
  BI-RADS 5 1 (1) 1 (1) 0 (0) (95% CI: 0%, 10.6%). The findings at
axillary screening US were negative in all
Note.—Data are numbers of examinations. Data in parentheses are numbers of women diagnosed with breast cancer within 1
year after screening US examination.
25 women with breast cancers detected
at supplemental screening US, including
two invasive cancers with axillary lymph
Table 3 node metastases found at surgery.

US Features of 33 Positive Axillary Findings at Screening US
Axillary Finding
Lymph node enlargement (n = 29)
  Unilateral (n = 23)
   BI-RADS 3 (n = 14)
   BI-RADS 4A (n = 9)
  Bilateral: BI-RADS 3 (n = 6)
Soft-tissue mass (unilateral): BI-RADS 4A (n = 3)
Mass in the accessory breast (unilateral):
  BI-RADS 4A (n = 1)
Note.—Unless otherwise indicated, data are medians, with ranges in parentheses. NA = not applicable.
* Data are numbers of cases.
axillary findings, but two women had a
history of breast cancer. Eleven of the
33 women had positive breast findings at
breast US (BI-RADS category 3 masses
in seven women and category 4A mass-
es in four women); none were diag-
nosed with breast cancer. The other 22
women showed positive findings only
in the axillae. Biopsy (n = 13) or fol-
low-up (n = 20) was recommended for
the 33 axillary findings. One lesion, uni-
lateral axillary lymphadenopathy, had
decreased in size at the time of biopsy
and was considered benign; therefore,
Screening Outcomes for Breast US with
and without Axillary Scanning
Lesion Size Round Cortical Loss of
(mm) Shape* Thickness (mm) Fatty Hilum* Baseline screening breast US with rou-
tine axillary scanning showed an AIR
of 15.2% (95% CI: 14.1%, 16.4%) and
enabled detection of an additional 1.5
11.4 (5.4–21.3) 4 4.0 (2.8–7.4) 4
(95% CI: 0.5, 3.3) cancers per 1000
15.3 (6.2–24.3) 0 5.3 (3.0–7.9) 2
examinations when the bilateral axillae
9.6 (6.6–22.9) 0 3.9 (3.0–5.4) 0
were routinely scanned (Table 4). The
16.2 (11.1–18.0) NA NA NA
PPV1 was 1.0% (six of 610 examina-
17.6 NA NA NA
tions), PPV2 was 4.6% (five of 109 ex-
aminations), and PPV3 was 6.0% (five of
83 examinations). Subsequent screening
breast US with routine axillary scanning
showed an AIR of 8.1% (95% CI: 7.5%,
8.7%) and a CDR of 2.2 cancers per
biopsy was cancelled. The woman had 1000 examinations (95% CI: 1.4, 3.4).
no evidence of malignancy after 22 When outcome measures of screening
months of clinical follow-up. Of the 12 US were calculated by using the breast
axillary lesions that were sampled for data to determine the hypothetical per-
biopsy, 11 were sampled with US-guided formance of screening breast US with-
16-gauge core needle biopsy and the out routine axillary scanning, the AIR
findings were reactive hyperplasia (n = 5) was 15.0% (95% CI: 13.9%, 16.2%) at
(Fig 2), lymphoid tissue with no tumor baseline screening and 7.9% (95% CI:
(n = 1), fibroadipose tissue only (n = 3), 7.4%, 8.5%) at subsequent screening.
and fibrocystic change (n = 2). With re- Incremental CDR was not changed with
gard to the finding of only fibroadipose axillary scanning for either baseline or
tissue at biopsy, the images and patho- subsequent screening examinations.
logic findings were considered to be Without axillary scanning, the PPV2
concordant in two women with axillary and PPV3 of screening US increased;

834 radiology.rsna.org  n Radiology: Volume 286: Number 3—March 2018

BREAST IMAGING: Effect of Routine Axillary US in Women with Negative Mammographic Findings Lee et al

Figure 2

Figure 2:  US images show axillary lymphadenopathy detected at supplemental screening US in 53-year-old woman with neg-
ative findings at mammography. (a, b) Two orthogonal views in left axilla at axillary screening US show 18.7-mm lymph node
with eccentric cortical thickening (cortical thickness, 5.3 mm; preserved fatty hilum). US-guided 16-gauge core needle biopsy
was performed in left axillary lymph node; result was reactive hyperplasia.

the PPV3 of baseline and subsequent Table 4

screening US increased from 6.0% (five
of 83 examinations) to 6.4% (five of 78 Outcomes of Screening Breast US with and without Axillary Scanning
examinations) and from 7.6% (13 of 170 Supplemental Screening US of Supplemental Screening US
examinations) to 7.9% (13 of 164 exam- Outcome Measure Breast and Axilla of Breast Alone
inations), respectively.
Baseline screening US
Characteristics of Cancers Detected with   AIR (%) 15.2 (610/4009) [14.1, 16.4] 15.0 (602/4009) [13.9, 16.2]
Screening US   CDR (‰) 1.5 (6/4009) [0.5, 3.3] 1.5 (6/4009) [0.5, 3.3]
 PPV1(%) 1.0 (6/610) 1.0 (6/602)
Of the 25 cancers detected with supple-
 PPV2 (%) 4.6 (5/109) 4.9 (5/103)
mental screening US, six (24.0%) were  PPV3 (%) 6.0 (5/83) 6.4 (5/78)
ductal carcinoma in situ and 19 (76.0%) Subsequent screening US
were invasive. Among the invasive   AIR (%) 8.1 (714/8835) [7.5, 8.7] 7.9 (700/8835) [7.4, 8.5]
cancers for which the size was known,   CDR (‰) 2.2 (19/8835) [1.4, 3.4] 2.2 (19/8835) [1.4, 3.4]
the mean size was 8 mm (median, 8  PPV1 (%) 2.7 (19/714) 2.7 (19/700)
mm; range, 1–15 mm). Of the 17 inva-  PPV2 (%) 7.0 (13/187) 7.2 (13/181)
sive cancers with known nodal status,  PPV3 (%) 7.6 (13/170) 7.9 (13/164)
15 (88.2%) had no axillary lymph node
Note.—Numbers in parentheses are raw data; numbers in brackets are 95% CIs.
metastasis and two (11.8%) had metas-
tases in one axillary lymph node. Three
interval cancers were identified in our
study population. Two of the three in- screening US enabled detection of 25 gathered from a single high-volume
terval cancers were invasive (mean size, mammographically occult breast can- screening center for 3 consecutive
15 mm), and none involved metastasis cers (19 invasive), and all of these can- years. In contrast, 33 positive findings
in axillary lymph nodes. cers were diagnosed as positive findings detected at axillary screening US in 33
in the breast. Axillary screening US women were all benign (95% CI for the
would have increased the additional rate of cancer among those with posi-
Discussion CDR if it had depicted axillary metas- tive axillary findings: 0%, 10.6%) in our
According to our study findings, rou- tases in women with occult primary study.
tine axillary scanning performed during breast tumors at screening breast US or The yield for the detection of ax-
supplemental screening breast US had primary cancers occurring in accessory illary metastasis is expected to be low
no effect on additional cancer detec- axillary breast tissue; however, there with axillary screening US because
tion, but instead increased the number were no such patients in our study screening US–detected breast cancers
of false-positive results. Supplemental population (n = 12  844), which was are mostly small, node-negative invasive

Radiology: Volume 286: Number 3—March 2018  n  radiology.rsna.org 835

BREAST IMAGING: Effect of Routine Axillary US in Women with Negative Mammographic Findings Lee et al
cancers (8–14). In our study,15 of 17
invasive cancers that had been detected
at supplemental screening US and had
undergone staging (88.2%) were node-
negative, similar to the findings of the
American College of Radiology Imag-
ing Network ACRIN 6666 trial (89%
[eight of nine]) and Japan Strategic
Anti-cancer Randomized Trial J-START
(86% [47 of 55]) (8,14). Even if metas-
tases are present in the axillary lymph
nodes, small metastases without en-
largement of the node or replacement
of the fatty hilum can have a normal ap-
pearance at US (26). The sensitivity of
US for the detection of axillary metas-
tases is known to be moderate (26.4%–
79.5%) when morphologic criteria are
used (27). In our study, two invasive
cancers were detected at supplemen-
tal screening US and axillary metasta-
ses were found at surgery; however,
neither showed suspicious findings at
axillary screening US. Primary cancers
occurring in accessory axillary breast
tissue are very rare and are less likely
to be diagnosed at screening US; pal-
pable symptoms in the axilla or abnor-
mal findings at mammography, includ-
ing calcifications in accessory axillary
breast tissue, are usually accompanied
by other findings (28).
Many studies (23,26,27) have eval-
uated the usefulness of US in assessing
axillary lymph nodes in patients with
known breast cancer; however, studies
evaluating the clinical importance of
axillary lymphadenopathy encountered
at axillary scanning during screening
breast US are few. In one such study,
Kim and Park (29) found that no ma-
lignancy was diagnosed in nonpalpable
axillary lymphadenopathies identi-
fied at breast US in patients without
known malignancy, despite frequent
manifestations of suspicious US fea-
tures including loss of the fatty hilum,
round shape, abnormal cortical thick-
ening, and marked hypoechogenicity.
Our results support their conclusion
that short-term follow-up imaging
rather than immediate biopsy can
be recommended for nonpalpable
lymphadenopathy in patients without
known malignancy. However, in the
study by Kim and Park, only 13 of 73
836 radiology.rsna.org  n Radiology: Volume 286: Number 3—March 2018
US examinations (17.8%) were per-
formed for screening purposes. Our
study included 12  844 supplemental
screening US examinations in women
with negative mammographic findings,
and the frequency of positive axillary
findings at screening US was 3.5 per
1000 baseline examinations and 2.2
per 1000 subsequent examinations. The
recall rate for the axilla was higher at
baseline screening US than at subse-
quent screening because comparisons
with axillary findings from previous US
examinations can reduce unnecessary
recall examinations. Without routine
axillary scanning, the AIR of screening
US decreased and the PPV2 and PPV3 of
screening US increased for both base-
line and subsequent examinations.
Our study had several limitations.
First, this study was a single-center
retrospective analysis, which limits
the generalizability of the study results.
Most of our study population (96.6%,
12 411 of 12 844 women) had no fam-
ily history of breast cancer in first-de-
gree relatives. Second, the number of
women in our study population with
abnormal axillary findings at screening
US was small (n = 33). The upper limit
of the rate of cancer among those with
positive axillary findings was 10.6%.
Further studies with larger patient
populations appear to be warranted.
Third, a positive screening US finding
was defined as a final BI-RADS assess-
ment category of 3 or higher based on
the report and did not strictly follow
the audit guidelines of the fifth edition
of BI-RADS (25). This was because our
study period preceded the publication
and implementation of the new edition
of BI-RADS, in which a screening US
examination is considered positive if
both standard (a single image of each
breast quadrant plus a retroareolar im-
age and, if a benign finding is present, a
single representative image of that find-
ing) and additional (diagnostic) images
are recorded, independent of the as-
sessment rendered. Screening US was
performed by radiologists in our study,
and additional diagnostic US images
were acquired for breast and axillary
lesions assessed as BI-RADS category
3 or higher.
In conclusion, routine axillary scan-
ning during supplemental screening
breast US did not provide additional
breast cancer detection, but rather
increased the number of false-positive
results leading to recall examinations
and biopsies. However, the conclu-
sions based on these findings must be
tempered by the low rate of positive
findings (CI for cancer detection after
axillary scanning: 0%, 10.6%). Our
study results suggest that additional
axillary US may not be necessary when
automated breast US or breast US
with a handheld device is performed in
women with negative findings at screen-
ing mammography.
Disclosures of Conflicts of Interest: S.H.L. dis-
closed no relevant relationships. A.Y. disclosed
no relevant relationships. M.J.J. disclosed no
relevant relationships. J.M.C. disclosed no rel-
evant relationships. N.C. disclosed no relevant
relationships. W.K.M. disclosed no relevant
relationships.
References
1. Tabár L, Vitak B, Chen TH, et al. Swedish
two-county trial: impact of mammographic
screening on breast cancer mortality during
3 decades. Radiology 2011;260(3):658–663.
2. Independent UK Panel on Breast Cancer
Screening. The benefits and harms of breast
cancer screening: an independent review.
Lancet 2012;380(9855):1778–1786.
3. Mandelson MT, Oestreicher N, Porter PL,
et al. Breast density as a predictor of mam-
mographic detection: comparison of interval-
and screen-detected cancers. J Natl Cancer
Inst 2000;92(13):1081–1087.
4. Kolb TM, Lichy J, Newhouse JH. Compari-
son of the performance of screening mam-
mography, physical examination, and breast
US and evaluation of factors that influence
them: an analysis of 27 825 patient evalua-
tions. Radiology 2002;225(1):165–175.
5. Harvey JA, Bovbjerg VE. Quantitative as-
sessment of mammographic breast density:
relationship with breast cancer risk. Radiol-
ogy 2004;230(1):29–41.
6. Boyd NF, Guo H, Martin LJ, et al. Mam-
mographic density and the risk and detec-
tion of breast cancer. N Engl J Med 2007;
356(3):227–236.
7. Berg WA. Current status of supplemental
screening in dense breasts. J Clin Oncol
DOI:10.1200/JCO.2015.65.8674. Published
online March 9, 2016. Accessed May 27,
2017.
BREAST IMAGING: Effect of Routine Axillary US in Women with Negative Mammographic Findings Lee et al
8. Berg WA, Blume JD, Cormack JB, et al.
Combined screening with ultrasound and
mammography vs mammography alone in
women at elevated risk of breast cancer.
JAMA 2008;299(18):2151–2163.
9. Corsetti V, Houssami N, Ferrari A, et al.
Breast screening with ultrasound in women
with mammography-negative dense breasts:
evidence on incremental cancer detection
and false positives, and associated cost. Eur
J Cancer 2008;44(4):539–544.
10. Youk JH, Kim EK, Kim MJ, Kwak JY, Son
EJ. Performance of hand-held whole-breast
ultrasound based on BI-RADS in women
with mammographically negative dense
breast. Eur Radiol 2011;21(4):667–675.
11. Hooley RJ, Greenberg KL, Stackhouse RM,
Geisel JL, Butler RS, Philpotts LE. Screening
US in patients with mammographically dense
breasts: initial experience with Connecticut
Public Act 09-41. Radiology 2012;265(1):
59–69.
12. Weigert J, Steenbergen S. The Connecti-
cut experiment: the role of ultrasound in
the screening of women with dense breasts.
Breast J 2012;18(6):517–522.
13. Chang JM, Koo HR, Moon WK. Radiologist-
performed hand-held ultrasound screening
at average risk of breast cancer: results
from a single health screening center. Acta
Radiol 2015;56(6):652–658.
14. Ohuchi N, Suzuki A, Sobue T, et al. Sensitiv-
ity and specificity of mammography and ad-
junctive ultrasonography to screen for breast
cancer in the Japan Strategic Anti-cancer
Randomized Trial (J-START): a randomised
controlled trial. Lancet 2016;387(10016):341–
348.
15. Cho N, Han W, Han BK, et al. Breast can-
cer screening with mammography plus ultra-
Radiology: Volume 286: Number 3—March 2018  n  radiology.rsna.org 837
sonography or magnetic resonance imaging
in women 50 years or younger at diagnosis
and treated with breast conservation ther-
apy. JAMA Oncol doi: 10.1001/jamaon-
col.2017.1256. Published online June 22, 2017.
Accessed May 27, 2017.
16. Brem RF, Tabár L, Duffy SW, et al. Assess-
ing improvement in detection of breast cancer
with three-dimensional automated breast US
in women with dense breast tissue: the So-
moInsight Study. Radiology 2015;274(3):663–
673.
17.
Wilczek B, Wilczek HE, Rasouliyan L,
Leifland K. Adding 3D automated breast ultra-
sound to mammography screening in women
with heterogeneously and extremely dense
breasts: report from a hospital-based, high-
volume, single-center breast cancer screening
program. Eur J Radiol 2016;85(9):1554–1563.
18. Ferré R, AlSharif S, Paré M, Kao E, Mesur-
olle B. Should the axilla be included in
screening US? Radiology 2015;274(2):623–
624.
19. Durand MA, Hooley RJ. Implementation
of whole-breast screening ultrasonography.
Radiol Clin North Am 2017;55(3):527–539.
20. Schelfout K, Kersschot E, Van Goethem M, et
al. Breast MR imaging in a patient with unilat-
eral axillary lymphadenopathy and unknown
primary malignancy. Eur Radiol 2003;13(9):
2128–2132.
21. Park YM, Park JS, Yoon HK, Yang WT.
Imaging-pathologic correlation of diseases in
the axilla. AJR Am J Roentgenol 2013;200(2):
W130–W142.
22. Sickles EA, D’Orsi CJ, Bassett LW, et al. ACR
BI-RADS mammography. In: ACR BI-RADS
atlas, breast imaging reporting and data
system. 5th ed. Reston, Va: American College
of Radiology, 2013.
23. Abe H, Schmidt RA, Kulkarni K, Sennett CA,
Mueller JS, Newstead GM. Axillary lymph
nodes suspicious for breast cancer metasta-
sis: sampling with US-guided 14-gauge core-
needle biopsy—clinical experience in 100
patients. Radiology 2009;250(1):41–49.
24. Mendelson EB, Bohm-Velez M, Berg WA,
et al. ACR BI-RADS ultrasound. In: ACR BI-
RADS atlas, breast imaging reporting and
data system. Reston, Va: American College
of Radiology, 2013.
25. Sickles EA, D’Orsi CJ. ACR BI-RADS
follow-up and outcome monitoring. In: ACR
BI-RADS atlas, breast imaging reporting
and data system. 5th ed. Reston, Va: Ameri-
can College of Radiology, 2013.
26. Ecanow JS, Abe H, Newstead GM, Ecanow
DB, Jeske JM. Axillary staging of breast
cancer: what the radiologist should know.
RadioGraphics 2013;33(6):1589–1612.
27. Alvarez S, Añorbe E, Alcorta P, López F,
Alonso I, Cortés J. Role of sonography in
the diagnosis of axillary lymph node metas-
tases in breast cancer: a systematic review.
AJR Am J Roentgenol 2006;186(5):1342–
1348.
28. Patel BK, Jafarian N, Abbott AM, Khazai
L, Lee MC. Imaging findings and manage-
ment of primary breast cancer in accessory
axillary breast tissue. Clin Breast Cancer
2015;15(4):e223–e229.
29. Kim SJ, Park YM. Clinical importance
and sonographic features of nonpalpable
axillary lymphadenopathy identified on
breast sonography in patients without ma-
lignancy. J Ultrasound Med 2015;34(12):
2193–2202.